The present invention relates to analyzing image data and, more particularly, to determining geometric parameters from a single image.
Geometric parameters, such as, for example, a length, a width, a height, and/or other parameters, of an object may be determined for a variety of applications and fields. For example, navigation systems or devices may use geometric parameter values of geographic objects to provide guidance and/or visual displays to end users. In one embodiment, a navigation system or device may provide realistic visual models of signs and other road features based on substantially accurate real-world geometric parameters of those signs and road features. Also, the navigation system or device may provide better route guidance, route calculation, safety features, and/or other navigation-related functions based on substantially accurate real-world geometric parameters, such as road lane widths.
Real-world geometric parameters, such as for road objects, may be determined by real-world measurements, such as by hired personnel. However, real-world measurements for multiple geographic features may be time consuming and costly if not impractical or improbable. Some systems may use multiple cameras to focus on a geographic feature, and correspondence calculations between the cameras are used to determine geometric parameters of the geographic feature. Also, multiple images of the same geographic feature may be used to determine geometric parameters, such as by triangulation. However, the use of multiple cameras and/or multiple images to estimate geometric parameters for a geographic object may involve increased complexity, time, and cost.